Patent Law

A Clustering of Patents

By Stephen Albainy-Jenei

PatentCluster, a free clustering based patent search engine, provides a different way to search patents.   PatentCluster was founded after observing that existing patent search engines did not make it easy to understand the context of the patent search results and did not articulate the “big picture” of the overall concept of the search.

Instead of giving the search user thousands of patent search results, PatentCluster search engine groups similar patents together into clusters. Clusters are meant to help you see your patent results by topic so you can zero in on exactly what you?re looking for or discover unexpected relationships between patents.

Using open source search and clustering algorithms to help discover diverse but relevant groups of semantically related patents, the clusters help you find results you may have missed or that were buried deep in the ranked list. You don’t waste time scrolling through page after page without a clear understanding of categories and relationships within the patent search results.

The site launched in January this year and is beta mode with a partial collection of the US patents. They plan to add the entire US issued and application patents in the coming months.  In a quick search of our own, our query for a particular drug turned up a single patent with no clusters (compared to 125 patents and applications found at the USPTO site) so you should plan to back up your search with using other methods.  Using the more generic term “enantiomers” produced 100 patents in 22 clusters.

Try your own search here.

Full post as published by Patent Baristas on January 26, 2009 (boomark / email).